Seasonality intensification and long-term winter cooling as a part of the Late Pliocene climate development

HAL Id: hal-00145040

https://hal.archives-ouvertes.fr/hal-00145040

Submitted on 25 Jun 2021

HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

Seasonality intensification and long-term winter cooling as a part of the Late Pliocene climate development

Stefan Klotz, Séverine Fauquette, Nathalie Combourieu-Nebout, Dieter Uhl, Jean-Pierre Suc, Volker Mosbrugger

To cite this version:

Stefan Klotz, Séverine Fauquette, Nathalie Combourieu-Nebout, Dieter Uhl, Jean-Pierre Suc, et al.. Seasonality intensification and long-term winter cooling as a part of the Late Pliocene climate development. Earth and Planetary Science Letters, Elsevier, 2006, 241, pp.174-187.

�10.1016/j.epsl.2005.10.005�. �hal-00145040�

Seasonality intensification and long-term winter cooling as a part of the Late Pliocene climate development

Stefan Klotz

^a,b,

*, Se´verine Fauquette

^c

, Nathalie Combourieu-Nebout

^d

, Dieter Uhl

^a,e

, Jean-Pierre Suc

^b

, Volker Mosbrugger

^a

aInstitut fu¨r Geowissenschaften, Universita¨t Tu¨bingen, Sigwartstr. 10, 72070 Tu¨bingen, Germany

bLaboratoire Pale´oEnvironnements et Pale´obioSphe`re (UMR 5125 CNRS), Universite´ Claude, Bernard- Lyon 1, 27-43 boulevard du 11 Novembre, 69622 Villeurbanne Cedex, France

cInstitut des Sciences de l’Evolution (UMR 5554 CNRS), Universite´ Montpellier II, Place Euge`ne Bataillon, 34095 Montpellier Cedex 05, France

dLaboratoire des Sciences du Climat et de l’Environnement (UMR 1572), L’Orme des Merisiers, Centre de Saclay, 91191 Gif-sur-Yvette Cedex, France

eLaboratory of Palaeobotany and Palynology, Department of Palaeoecology, University of Utrecht, Budapestlaan 4, 3584 CD Utrecht, The Netherlands

Received 16 March 2005; received in revised form 15 September 2005; accepted 5 October 2005 Available online 21 November 2005

Editor: H. Elderfield

Abstract

A mutual climatic range method is applied to the Mediterranean marine pollen record of Semaforo (Vrica section, Calabria, Italy) covering the period from ~2.46 Ma to ~2.11 Ma. The method yields detailed information on summer, annual and winter temperatures and on precipitation during the nine obliquity and precession-controlled

d

glacial

T

periods (marine isotope stages 96 to 80) and eight

d

interglacial

T

periods (marine isotope stages 95 to 81) characterising this time interval. The reconstruction reveals higher temperatures of at least 2.8

8C in mean annual and 2.28C in winter temperatures, and 500 mm in precipitation during the d

interglacials

T

as compared to the present-day climate in the study area. During the

d

glacials

T

, temperatures are generally lower as compared to the present-day climate in the region, but precipitation is equivalent. Along the consecutive

d

interglacials

T

, a trend toward a reduction in annual and winter temperatures by more than 2.3

8C, and toward a higher seasonality is observed. Along the

consecutive

d

glacials

T

, a trend toward a strong reduction in all temperature parameters of at least 1.6

8C is reconstructed. Climatic

amplitudes of

d

interglacial–glacial

T

transitions increase from the older to the younger cycles for summer and annual temperatures.

The cross-spectral analyses suggest obliquity related warm/humid–cold/dry

d

interglacial–glacial

T

cycles which are superimposed by precession related warm/dry– cold/humid cycles. A time displacement in the development of temperatures and precipitation is indicated for the obliquity band by temperatures generally leading precipitation change at ~4 kyr, and on the precession band of

~9.6 kyr in maximum.

D

2005 Elsevier B.V.

Keywords:Late Pliocene; obliquity; precession; palaeoclimate; pollen

1. Introduction

Research into the Late Pliocene climate is of espe- cial interest as it contributes to the understanding of

0012-821X D2005 Elsevier B.V.

doi:10.1016/j.epsl.2005.10.005

* Corresponding author. Institut fu¨r Geowissenschaften, Universita¨t Tu¨bingen, Sigwartstr. 10, 72070 Tu¨bingen Germany. Tel.: +49 7071 2973081; fax: +49 7071 295217.

E-mail address:stefan.klotz@uni-tuebingen.de (S. Klotz).

Earth and Planetary Science Letters 241 (2006) 174 – 187

www.elsevier.com/locate/epsl

Open access under CC BY-NC-ND license.

Open access under CC BY-NC-ND license.

the nature and the processes of global cooling that evolved since the Miocene and intensified after ~2.8 million years ago, exhibiting an alternation of warm (dinterglacialT) and cool-temperate (dglacialT) periods (e.g., [1-6]). The Late Pliocene palaeoclimate response of the Mediterranean is well documented by marine proxy data [7–11] and few vegetation records [6,12, 13]. The marine data revealed the strong influence of the ~41 kyr obliquity periodicity dominantly affecting high latitude climate and leaving its signature in less pronounced d interglacial–glacial T cycles [14], and the

~23 kyr precession periodicity mainly influencing lower latitude monsoonal activity [11]. Especially, it has been shown that Sea Surface Temperature changes in the Mediterranean were obliquity-controlled and caused by dinterglacial–glacialT variability, whereas Sea Surface Productivity and Salinity changes were precession-controlled and associated with sapropel for- mation [9,15]. The latter were interpreted as reflecting the influence of the African monsoon. In addition, in a high-resolution marine pollen record of the Semaforo section (Calabria, Italy) covering MIS 96 to 89 (~2.3 to ~2.17 Ma) [5,16], the alternation between warm humid d interglacials T (obliquity maxima) and cold- temperate dry d glacials T (obliquity minima) was docu- mented, the cycles of which were assumed to be superimposed by relatively warm humid conditions at times of precession minima and cold dry conditions during precession maxima [15]. So far, however, quantification of terrestrial temperature and precipita- tion changes on orbital scales are rather sparse for the Late Pliocene, especially in the Mediterranean. The main reason may be first the very few available (relative high-resolution) terrestrial palaeoclimate archives, such as pollen records. In addition, recon- struction methods are required which do not consider modern analogue floras since for older periods (i.e.

pre Mid-Pleistocene) frequently floras exist which have no present-day equivalents. Therefore, potentially mutual climatic range methods are applicable (e.g., [17-21]) which determine the ranges of the climatic tolerances of fossil floras by means of the mutual present-day ranges of the nearest living relatives (NLR) of the plants represented in the fossil assem- blages. However, so far the climatic resolution is often low with these methods and the databases of the climate requirements of the present-day nearest living relatives of fossil plants are often poor. Hence, we performed a sophisticated mutual climatic range meth- od to provide quantitative temperature and precipita- tion estimates for the extended data set of the Semaforo pollen record [22] covering the interval of

MIS 96 to 80, and to analyse the reconstruction results for palaeoclimatic characteristics. Specifically, we ad- dress the following questions: (1) which main climatic trends related to atmospherical changes characterise the successive Late Pliocene cold and warm phases, and (2) how do the vegetation pattern and the recon- structed climate parameters correspond to the obliquity and precession signals?

2. The marine record of Semaforo and its chronology

The Semaforo section contains a marine record exposed South of Crotone/Italy (398N, 16842VE) (Fig. 1). The record consists of 180 m of marine clays, comprising 8 sapropel layers (layers 222 to 204 after Hilgen [14]) and 3 volcanic ash beds (near sapropel layers 222, 208, 206). The time stratigraphic frame of the Semaforo record is well established by astronomically tuned sapropel chronology used for the standard geologic time scale, as well as the standard isotope stratigraphy [15,16,23], suggesting to cover the time span from ~2.46 to ~2.11 Ma (Fig. 2).

With respect to the international agreement placing the Pliocene–Pleistocene boundary near the top of the Olduvai subchron at 1.086 Ma as defined in the overlying Vrica section/Italy [23,24], the Semaforo section belongs to the Late Pliocene, the Gelasian Stage. Palynologically, the record is represented in

Fig. 1. Site map of the study area with the sections Semaforo and Monte Singa.

165 pollen spectra with 110 taxa identified, sampled along 175 m of sediments on intervals between 0.7 and 2 m [16]. For each sample which has been processed according to the standard HF method [25], a number between 200 and 1400 pollen grains have been counted, and abundances have been plotted into a pollen diagram (Fig. 3). The pollen floras sensitively reflect the palaeoclimate development of dinterglacial–

glacialT cycles during the Late Pliocene by the repet- itive alteration of main vegetation units. These units, which are grouped with respect to the phytogeograph- ical affinities of plants, repeatedly change from decid- uous forest (main abundant taxa are Quercus, Acer, Ilex, Carpinus, Buxus, Eucommia, Carya, Juglans, Pterocarya, Ulmus) to subtropical humid forest (Cath- aya, Engelhardia, Magnolia, Nyssa, Symplocos , Tax- odiaceae), to altitudinal forest (Tsuga, Cedrus, Picea, Abies), and to open herbaceous formation (Chenopo- diaceae, Poaceae, Artemisia, Ephedra, Asteraceae).

Another vegetation unit is represented by Mediterra- nean ecosystems (Olea, Ceratonia, Pistacia, Quercus ilex, Cistus, Phillyrea, Ligustrum), which irregularly follows deciduous forest or subtropical humid forest (Fig. 4). From the base to the top of the fossil pollen record, subtropical humid forest elements (especially Taxodiaceae) are increasingly replaced by deciduous forest elements (especially Quercus), until the subtrop- ical ecosystems finally disappear in the Mediterranean at ~1.2 Ma [5,22]. Simultaneously, a general increase in altitudinal forest elements and Artemisia as an indicator for open vegetation is observed, but tree pollen always dominate over herbaceous pollen. Climatically, this pattern has been interpreted to suggest (1) that dinterglacialT temperature maxima may decrease during

the course of the Late Pliocene, (2) that aridity or seasonality increases, (3) that dglacialT times experi- enced enhanced aridity, and (4) that precipitation decreases later than temperatures during dinterglacial–

glacial T transitions [16,22].

3. Method used for palaeoclimate reconstruction For a sophisticated interpretation of palaeoclimate signals, the bprobability mutual climatic spheresQ (PCS) described in detail by Klotz [20] and Klotz et al. [21,28] is applied, taking into account 60 out of 110 taxa identified in the fossil pollen floras for reconstruction (Table 1). The principle of PCS is based on the calculation of palaeoclimate ranges for a fossil flora by means of mutual 2-dimensional cli- matic spheres of the NLR of the individual plants which participate in the fossil flora. The 2-dimensional spheres representing the present-day climate require- ments of the NLR are derived from the correlation of present-day climate data with digitised potential dis- tribution maps of plants. Within the mutual 2-dimen- sional climatic sphere of a fossil flora, probability intervals are calculated for the individual climate para- meters. For explanation, we refer to mean annual temperature (MAT) as an example. First, the range of MAT as inferred from the mutual climatic sphere of the fossil flora is compared to those MAT ranges which have been calculated for 9555 present-day floras, and for which present-day climate data are available. It can be observed that the actual MAT values of those present-day floras which have similar ranges as the fossil flora accumulate within a smaller interval of the total range, indicating the climatic

Fig. 2. Stratigraphic column with chronology, polarity zones and lithology[16], MPRS coding[14], pollen samples and biostratigraphy[16].

S. Klotz et al. / Earth and Planetary Science Letters 241 (2006) 174–187 176

preference of the floras. The interval of preference is then interpreted as the probability interval of MAT.

For representation of the reconstructions, we use the

centres of the probability intervals. The quality of PCS has been tested on the base of a multitude of present- day floras [20,21], yielding information on the large

Fig. 3. Pollen diagram of the Semaforo record according to Combourieu-Nebout[22]with thickness, lithology[16](same caption as inFig. 2) and MPRS coding[14]. Percentages ofPinusare calculated on the total pollen and spores sum, percentages of other taxa and AP, NAP are based on the total sum excludingPinus.

agreement between reconstructed and actual climate values (for which the correlation coefficients and mean aver age error are 0.95, 1.1 8C for summer

temperatures, 0.95, 1.7 8C for winter temperatures, 0.95, 1.1 8C for mean annual temperature, and 0.86, 100 mm for mean annual precipitation, respectively).

Fig. 4. Generalised pollen diagram of the Semaforo record represented by the percentages of the main vegetation units deciduous forest, subtropical humid forest (most abundant taxa are Taxodiaceae andCathaya), Mediterranean ecosystems, altitudinal forest, and open herbaceous formation (most abundant taxa are Amaranthaceae–Chenopodiaceae andArtemisia). Abundance of subtropical humid + deciduous forest reflects a qualitative land temperature (LT) index according to Lourens[15]. Correlation of the Semaforo pollen record with relevant marine indices: The LT index, Sea Surface Temperatures (SST) and planktonicd¹⁸O data from the Singa section (3883VN, 1683VW, ~100 km South South–West of Semaforo)[9], and benthicd¹⁸O record from DSDP 607 in the North Atlantic[1]are tied in the lower section according to Lourens[15]and Lourens et al.[9]

identifying marine isotope stages (MIS) 96 to 90. This scheme is continued for the upper section. Correlation is made also with the benthicd¹⁸O record from eastern equatorial Pacific site ODP 849[26]and with orbital signatures[27].

S. Klotz et al. / Earth and Planetary Science Letters 241 (2006) 174–187 178

Specifically, the high correlation coefficients indicate that PCS is able to reliably reproduce trends.

4. Late Pliocene climate

The application of PCS on the Semaforo pollen record allows a detailed view on the Late Pliocene climate development in the Mediterranean, revealing the succession of eight d interglacials T and nine dglacialsT (Fig. 5). Changes in temperature parameters are mostly correlated with each other (R

²

N 0.67, p b 0.01), whereas temperature parameters and precip- itation are autocorrelated (temperatures leading by ~4 kyr, R

²

N 0.47, p b 0.01). As compared to temperature parameters, precipitation shows larger relative fluctua- tions during dinterglacialsT. Generally, average oscilla- tions in the transition between dinterglacialsT and dglacialsT are higher in mean temperatures of the coldest month (MTC, 7.3 8C) than in mean annual temperatures (MAT, 6.6 8C) and in mean temperatures of the coldest month (MTW, 5.3 8C); mean annual

precipitation (MAP) shows an average range of 540 mm. Maximum summer temperatures during the con- secutive dinterglacialsT vary between 23.8 and 25.1 8C, annual temperatures between 13.2 and 15.5 8C, winter temperatures between 8.0 and 10 8C, and annual precipitation between 985 and 1444 mm. Lowest sea- sonality and highest oceanicity is revealed for MIS 95 (MTW-MTC is 14.6 8C, MAP is 1444 mm), highest continentality for MIS 81 (MTW-MTC is 16.5 8C). As compared to the present-day climate at Semaforo, during MIS 95, MTW is 2.4 8C, MAT is 1.9 8C, MTC is 3.3 8C, and MAP is 600 mm higher (note that the reconstructions integrates also the climatic signals from high altitude forest elements). Climate conditions during MIS 95 may be most similar to those at present of some regions in Central South China (~248 to 278N, ~1028 to 1078E) and South- Eastern South America (~328 to 368S, ~548 to 618E), both regions which are strongly influenced by season- al reversal (monsoonal) winds. Minimum summer temperatures during the consecutive dglacialsT vary

Table 1

Pollen taxa identified in the Semaforo record, grouped according to the phytogeographical affinites of the plants[5]

Deciduous forest Subtropical humid forest Centaurea Non classified taxa

!Acer !Cathaya Compositae Asteroideae Abietaceae

!Alnus Distylium Compositae Cichorioideae Araliaceae

!Betula !Engelhardia Brassicaceae !Armeria

!Buxus Magnolia !Cyperaceae Boraginaceae

!Carpinus !Myrica !Ephedra Caprifoliaceae

!Carya !Nyssa Lygeum Convolvulaceae

!Celtis Palmae !Poaceae Euphorbia

!Cornus Sapotaceae Phlomis Euphorbiaceae

!Corylus Sciadopitys !Plantago Gentianaceae

Elaeagnus !Simplocos !Rumex Geranium

!Ericaceae !Taxodiaceae Apiaceae !Helianthemum

!Eucommia Taxodium Meditteranean sclerophyllous forest !Knautia

!Fagus Altitudinal Coniferous forest Ceratonia Lamiaceae

!Hedera !Abies Cistus Fabaceae

!Ilex !Cedrus Citrus Lilium

!Juglans !Picea Cupressaceae Liliaceae

Liquidambar Tsuga !Fraxinus Malvaceae

!Lonicera Open vegetation Jasminus Nolina

!Ostrya Amaranthaceae !Ligustrum Nympheaceae

Parrotia !Chenopodiaceae Myrtaceae Fabaceae Papilionioideae

!Platanus !Artemisia !Olea Plumbaginaceae

!Populus Caryophyllaceae !Phillyrea !Potamogeton

!Pterocarya !Pinus Ranunculaceae

!Quercus !Pistacia Rosaceae

!Salix !Quercus ilex type Rhus

!Sambucus !Rhamnus Solanaceae

Taxaceae !Sparganium

!Tilia Tymeleaceae

!Ulmus Typha

!Vitis Valerianaceae

!Zelkova

Dots indicate those taxa which are used for palaeoclimate reconstruction.

between 17.5 and 20.5 8C, annual temperatures be- tween 7.2 and 11.2 8C, winter temperatures between 1.4 and 3 8C, and annual precipitation between 780 and 990 mm. Lowest summer, annual, and winter temperatures are found for MIS 80 (MTW-MTC is 16.1 8C). As compared to the present-day climate at Semaforo, during MIS 80, MTW is 4.7 8C, MAT is 6.4 8C, MTC is 5.3 8C, and MAP is 70 mm lower.

Climate conditions during MIS 80 may be most sim-

ilar to those at present of some regions in South France (~448 to 478N, ~28 to 58E) and Central West Italy (~428 to 448N, ~128 to 138E). The reconstructed values for the dinterglacialsT and dglacialsT are reason- ably in agreement with estimates for the Mediterra- nean pollen record of Garraf 1, northeastern Spain [19,29], covering the time window ~5.3 to ~1.75 Ma in 48 pollen samples (10 samples during ~2.5 to ~2 Ma). The deviations between the reconstructed aver-

Fig. 5. Climate reconstruction with the PCS method for the Semaforo pollen record, represented by the centres of the mutual climatic ranges for the parameters mean temperature of the warmest month (MTW), mean annual temperature (MAT), mean temperature of the coldest month (MTC) and mean annual precipitation (MAP). Note that although the centres of the probability intervals of PCS point to specific values, it is emphasised that the method basically provides intervals reflecting the ranges of actual palaeoclimate data. Standard deviation (r) from the centres is 18C for MTW, 0.678C for MAT, 0.98C for MTC, and 166 mm for MAP. Modern climate data at Semaforo (MTW 22.28C, MAT 13.68C, MTC 6.78C and MAP 850 mm) are shown on figure, represented by dotted vertical lines. The climate record is shown together with the planktonic d¹⁸O record, marine isotope stages (MIS), and sea surface temperatures from the Singa section [9], with the benthicd¹⁸O record from DSDP 607 [1] and eastern equatorial Pacific site ODP 849 [26], with obliquity and with precession modulated 658N June insolation[27].

S. Klotz et al. / Earth and Planetary Science Letters 241 (2006) 174–187 180

age climate values at Garraf 1 and Semaforo largely correspond to the present-day climate gradient be- tween these localities (Table 2). Hence, our recon- structions for that period reflect broad regional climatic patterns.

5. Discussion

The Semaforo pollen record provides an excellent opportunity for the study of the Mediterranean vegeta- tion and climate response on orbital forcing during the Late Pliocene within the time window ~2.46 to ~2.11 Ma. The cyclic vegetation changes recorded in the Semaforo sequence have been proposed to result main- ly from variations in temperature and moisture [22,30], affecting shifts in vegetation belts over the Sila Massif 30 km North of Semaforo (palaeoaltitude more than 1500 m [31]) which represents the main pollen source area [32]. In this regard, the change from subtropical forests to open herbaceous formation would reflect the shift from warm and humid dinterglacialsT to cold and/

or dry dglacialsT. The potential influence of an uplift of the Sila Massif (today 2000 m in altitude) on exposing larger areas for open vegetation (or altitudinal forest) which could influence the palaeoclimate reconstruction, can mostly be disregarded since (1) a constant uplift during the past ~2.46 Ma would result in an uplift of

~70 m during the period covered by the Semaforo record, and (2) there is evidence that the uplift took place after the deposition of marine terraces during or after the middle Pleistocene since these terraces have protected the Semaforo section after its outcrop. Using the PCS method for interpretation of the palaeoclimatic signals reflected by the vegetation, detailed in forma- tion is provided for temperature and precipitation change, suggesting general trends along the successive dinterglacialsT and dglacialsT.

5.1. Relationship to orbital parameters and d

¹⁸

O data The spectral analyses on selected vegetation units and on groupings reveal that the vegetation largely responded to the orbital forcing of the ~41 and ~23 kyr periodicities, whereas the ~100 kyr component is weak (Fig. 6a). Interestingly, the obliquity signal is best reflected by subtropical humid forest, altitudinal forest and open herbaceous vegetation, the first of which is also showing a precession component. In contrast, from all the vegetation units, Mediterranean ecosystems and deciduous forest show strongest response to the preces- sion band. With view on groupings of vegetation units (subtropical humid forest + deciduous forest, and sub- tropical humid forest + altitudinal forest), combined effects are observed in that the spectral patterns of these groups are largely determined by the interference of the spectral patterns of the individual vegetation units. Concerning the reconstructed climatic parameters which interpret the vegetation signals, the spectral anal- ysis suggests a strong control of the ~41 kyr periodicity on variations in temperatures and notably also in pre- cipitation. In contrast, the ~23 kyr periodicity operates stronger on precipitation changes than on temperatures.

(Fig. 6b).

5.1.1. Obliquity forcing

The obliquity controlled variations in the Semaforo record, and in the d

¹⁸

O data and Sea Surface Tempera- tures from the Singa section can easily be correlated with the d

¹⁸

O records of sites DSDP 607 and ODP 849 (see Fig. 5), therefore arguing to reflect dinterglacial–

glacialT cycles bound to ice volume [9]. Based on cross- spectral analysis, the calculated phase relationships re- veal that maximum obliquity is closely related to the abundance of warm humid vegetation elements, and to the maxima of pollen derived climate parameters and Singa SST (Fig. 7). For exact phase angles and coher- ency see Table 3. As a specific pattern, precipitation lags temperature parameters which may be easily seen by comparison of winter temperatures and precipitation for MIS 95, 91, 89, 85 and 83 (Fig. 6). This finding is in agreement with the palynological evidence [16], sug- gesting warm/dry conditions during the onset of dinterglacialsT by the rise of deciduous forest elements, and cold/humid conditions during the onset of dglacialsT by the rise of altitudinal forest elements. If calculating the mean time displacement in the development espe- cially between winter temperatures and precipitation along the dglacial–interglacialT transitions MIS 96/95 to 82/81 (time of each first sample is taken indepen- dently for winter temperatures and precipitation, when

Table 2

Present-day climate conditions at the localities of Semaforo and Garraf 1 (41810VN, 2801VE) with respect to the mean temperature of the warmest (MTW) and coldest month (MTC) in 8C, and mean annual precipitation (MAP) in mm

MTW MTC MAP

Present-day Semaforo 22.2 6.7 850

Garraf 1 21.4 6.5 690

gradient S-G 0.8 0.2 160

dInterglacialsTmaxima Semaforo 25.0 9.5 1370

Garraf 1 25.7 9.2 1100

dGlacialsTminima Semaforo 19.8 2.3 830

Garraf 1 22.0 1.0 600

Average values are calculated for the maxima of thedinterglacialsTand minima of thedglacialsTfor both records.

Fig. 6. Power spectra using fast Fourier transformation for different filtered parameters (5% tapered, 7 band Tukey-window). a) Left panel with power spectra for the vegetation units subtropical humid forest, deciduous forest, Mediterranean ecosystems, altitudinal forest and open herbaceous formation, and on the groups. subtropical + deciduous forest and subtropical humid + deciduous + altitudinal forest. b) Right panel with power spectra for the climate parameters mean temperatures of the warmest (MTW) and coldest month (MTC), for mean annual precipitation (MAP), and planktonicd¹⁸O data from the Singa section for comparison[9]. Confidence intervals (ci) for the parameters shown on figure.

Fig. 7. Cross-spectral analyses using fast Fourier transformation for different filtered parameters (5% tapered, 7 band Tukey-window). Left panel: wheel phase spectrum including different vegetation units, groupings of them, and climate parameter related to obliquity. Right panel: the same for precession.

S. Klotz et al. / Earth and Planetary Science Letters 241 (2006) 174–187 182

8 8C and 1360 mm levels defining the dinterglacialsT b plateaus Q are exceeded), a value of ~4.0 kyr is obtained (standard deviation ~3.2 kyr, mean time resolution during these transitions is ~1.7 kyr). For the mean time displacement calculated on the consecutive dinterglacial–glacialT transitions MIS 95/94 to 81/80 (when parameter values are below 8 8C and 1360 mm levels), a value of ~3.3 kyr is obtained (standard devi- ation ~2.0 kyr, mean time resolution during these tran- sitions is ~1.2 kyr). These time lags correspond also to the results provided by a cross-spectral analysis on MTC and MAP, yielding a general phase relationship between these parameters on the obliquity band of

~3.9 F 0.4 kyr (coherency of 0.9). The time lag between precipitation and temperatures can mainly be explained by means of the SST. During the onset of dinterglacialsT sea surface evaporation influencing precipitation is low due to still cool SST, whereas during the onset of d glacials T still considerable SST allow higher evapora- tion and precipitation, also supported by lower satura- tion vapor pressure of the air.

5.1.2. Precession forcing

Since the precession component reflected by the Mediterranean records is almost absent in the d

¹⁸

O data of sites DSDP 607 and ODP 849, it can not be interpreted in terms of ice volume changes but was linked to increased Sea Surface Temperatures (SST) at times of sapropel formation related to variations in freshwater input [9,15]. In contrast to the cyclic patterns related to obliquity, the calculated phase relationships

on the precession band indicate that precipitation is considerably leading minimum precession (maximum insolation), and is more phase related to altitudinal forest and more in anti-phase with SST, MTW and MTC.

According to the cross-spectral analyses, the maxi- mum phase lag of MAP behind MTC is ~9.6 kyr (coherency of 0.73). As a generalised result, the pre- cession related cycles are determined by the alternation of warm/dry–cold/humid phases. This view deviates from the classical (qualitative) pattern of precession related warm/humid–cold/dry cycles [15]. Our results especially suggest that precession related maximum precipitation mainly corresponds to times of generally low temperatures, which may be explained by (1) in- creased advection and a presumably more northwestern cyclonic wind regime, and/or (2) enhanced relative humidity due to lower saturation vapor pressure of the colder air.

As a general scheme of the palaeoclimatic develop- ment during the period covered by the Semaforo record, it has been considered that the obliquity related warm/

humid–cold/dry cycles are superimposed by the preces- sion related warm/dry–cold/humid cycles.

5.2. Long-term climatic trends

The reconstructed climate data are analysed for gen- eral trends calculating linear regressions for climate parameters along the eight dinterglacialsT and nine dglacialsT, that is for the development of dinterglacialT maxima, dglacialT minima, and dinterglacial–glacialT amplitudes. Unlike the various marine palaeoclimate indices which reveal no obvious characteristics along MIS 97 to MIS 80, the vegetation based climatic ana- lysis suggests different long-term trends of atmospheric change in the Mediterranean.

5.2.1. Annual and winter cooling along consecutive d interglacials T

The coldest dinterglacialT MIS 81 is even warmer in summer temperatures of ~3 8C, in winter temperatures of ~2 8C, and with higher precipitation of ~500 mm than the modern climate conditions at Semaforo. Re- ferring to the analysis of the dinterglacialsT maxima along MIS 95 to 81 (total period of ~280 kyr), strong trends of decreasing annual temperatures of ~2.8 8C and winter temperatures of ~2.3 8C become obvious (Fig. 8a). Due to similar summer temperatures, season- ality is increasing during the consecutive dinterglacialsT by ~2.3 8C. Although the general decrease of subtrop- ical humid elements and their replacement by decidu-

Table 3

Phase angles in degree (deg.) for the different parameters against obliquity and precession, and coherency (k) (at 95% confidence level)

Parameter ~41 kyr ~23 kyr

deg. k deg. k

Subtropical 12F8 0.87 173F24 0.16

Deciduous 18F10 0.62 170F3 0.62

Mediterranean 45F7 0.56 81F3 0.76

Altitudinal 167F3 0.91 104F2 0.66

Open 156F20 0.47 173F3 0.37

Sub. + Dec. 12F4 0.76 175F10 0.49

Sub. + Dec. + Alt. 12F6 0.73 162F9 0.61

SST Singa 52F4 0.41 110F1 0.97

MTC 30F1 0.69 88F2 0.85

MTW 25F5 0.66 91F4 0.91

MAP 37F13 0.59 41F22 0.96

Parameters include subtropical humid forest (Subtropical, Sub.), de- ciduous forest (Deciduous, Dec.), Mediterranean ecosystems (Medi- terranean), altitudinal forest (Altitudinal, Alt.), open herbacaeous formation (Open), Sea Surface Temperatures of the Singa record (SST Singa), mean temperatures of the coldest (MTC) and warmest month (MTW), and mean annual precipitation (MAP).

ous forest elements [22], the annual precipitation shows no trend. This can be explained by the following: (1) the occurrence of subtropical humid elements during all dinterglacialsT (abundance N 6%) indicate warm and (year-long) humid climate conditions at all [5,19,33–

36]. (2) The subtropical humid elements are more sensitive to temperature changes rather than precipita- tion changes, as can be interpreted from the Semaforo and Vrica composite section (Crotone, Italy) spanning the period from ~2.4 to ~1.2 Ma. There, Taxodiaceae

are gradually replaced by Cathaya between ~2.4 and

~2.0 Ma, and Cathaya is replaced by Tsuga after ~1.4 Ma [5,30]. Today, however, these taxa can be found at different altitudes with Cathaya and Tsuga at higher levels than Taxodiaceae, but all requiring similar levels of high humidity . The finding of decreasing winter temperatures is further in agreement with estimates derived from the northern German record of Lieth (53843VN, 9838VE) covering MIS 97 to 81 in about 49 pollen samples, based on the same reconstruction meth- od [37]. It has been proposed for Lieth a gradual decrease in winter temperatures of ~5 8C during the consecutive dinterglacialsT, and also an increase in sum- mer temperatures of ~1 8C. Regardless of the difference in the reconstructed absolute decreases in winter tem- peratures at Semaforo and Lieth, the comparison emphasises the over-regional character of the deter- mined trend along the consecutive dinterglacialsT during the period considered.

5.2.2. General cooling along consecutive dglacialsT Strong cooling trends are also observed along d glacials T MIS 96 to 80 (period of ~300 kyr) for all temperature parameters, with decreasing summer tem- peratures of 2.8 8C, mean annual temperatures of 3.2 8C, and winter temperatures of 1.6 8C (Fig. 8b). This finding coincides with the general increase of altitudinal forest elements and open vegetation elements, but does not reveal enhanced aridity. Considering the different marine indices, only the d

¹⁸

O data from the Singa record exhibits a weak trend of increasing values along MIS 96 to MIS 78 (+ 1.3 x , R

²

= 0.4). Our results are also in agreement with the pollen based reconstruc- tions from the record of Lieth, where a decrease in winter temperatures of ~5.5 8C along MIS 96 to 80 is suggested [37]. The intensified cooling during the con- secutive obliquity dominated dglacialsT may correspond to the strong imprint of the obliquity signal caused by the control meridional temperature gradients exert on the poleward transport of moisture, allowing to progres- sively build-up Northern Hemisphere ice sheets [38]. It may also be related to intensified atmospheric circula- tion [39], or to enhanced ice-rafting as evidenced from records in the Norwegian Sea and North Atlantic during dglacialsT and dinterglacial–glacialT transitions [40].

5.2.3. Increasing amplitudes of consecutive dinterglacial–glacialT transitions

Calculating the maximum climatic amplitudes be- tween consecutive dinterglacialsT (maximum values are considered) and dglacialsT (minimum values are consid- ered), another trend suggests increasing summer tem-

Fig. 8. Linear climatic trends along the consecutive dinterglacialsT MIS 97 to 81 anddglacialsTMIS 96 to 80 in the Semaforo record for the parameters mean temperature of the warmest month (MTW), mean annual temperature (MAT) and mean temperature of the coldest month (MTC), related to timescale. (a) Maximum values during the dinterglacialsT. (b) Minimum values during thedglacialsT. Coefficients of determination (at significance levelpb0.01) for regressions, and resulting total changes between the lower and upper part of the record are shown on figure. (c) Maximumdinterglacial–glacialTamplitudes (difference betweendinterglacialTmaximum and succeedingdglacialT minimum values).

S. Klotz et al. / Earth and Planetary Science Letters 241 (2006) 174–187 184

perature amplitudes of ~3.2 8C and mean annual tem- peratures of ~1.9 8C along MIS 96/95 to 81/80 (Fig.

8c). For comparison, the analysis on the d

¹⁸

O data from the Singa record reveals a trend suggesting higher amplitudes along MIS 97/96 to MIS 79/78 (+ 2.4 x , R

²

= 0.6), which is not obvious from the d

¹⁸

O record at DSDP 607 in the North Atlantic. It is especially sug- gested that changes in global ice volume associated with reductions in the Mediterranean surface-water sa- linity (resulting from changes in the rates of evapora- tion/precipitation and from fluvial runoff), led to higher dinterglacial–glacialT amplitudes in the region [41,42].

As compared to the amplitude of winter temperature change of more than 18 8C during the last dinterglacial–

glacialT transition (Eemian–Weichselian) as recon- structed from the pollen record of Lago Grande di Monticchio/Southern Italy [43], the Late Pliocene tran- sitions show less than half in amplitude.

6. Conclusions

The quantitative climate analysis of the Mediterra- nean pollen record of Semaforo/Southern Italy provides new and detailed information on the climate develop- ment along eight Late Pliocene d interglacials T and nine dglacialsT during the time window 2.46–2.11 Ma. Mainly forced by the obliquity and the precession signal modu- lating 658N June insolation, the analysis of climate development reveals specific characteristics of change.

(1) Summer, mean annual and winter temperatures, as well as precipitation are considerably higher during the Late Pliocene subtropical humid dinterglacialsT as compared to present-day climate conditions in the study area. Temperatures during dglacialT periods are lower, but precipitation is similar than at present.

(2) Relying on the timescale of the Semaforo record, the spectral analyses on vegetation units and on reconstructed climate parameters reveals the con- siderable imprint of the ~41 kyr obliquity and ~23 kyr precession periodicity. Generalised, it is evidenced the alternation of warm/humid–cold/

dry dinterglacial–glacialT cycles on the obliquity band, which are superimposed by warm/dry–cold/

humid cycles on the precession band. Whereas the obliquity bound climatic cycles in the Mediterra- nean largely follows the global change in ice vol- ume, the precession related cycles may be best explained by the influence of warm/dry southern anticyclonic wind circulation alternating with a cold/humid northwestern cyclonic wind regime.

(3) A time lag of ~4 kyr is observed in the development of temperatures and precipitation during the obliq- uity controlled dglacial–interglacialT cycles, and of

~9.6 kyr in maximum during the precession related cycles, with precipitation lagging temperature parameters. Transitional phases are thus charac- terised by warm/dry or cold/humid conditions.

(4) Along the consecutive dinterglacialsT no precipi- tation trend change is yielded according to the abundance of taxa requiring a high amount of year-long humidity.

(5) A strong trend in the climate development of over- regional scale along the consecutive Late Pliocene dinterglacialsT is indicated by the reduction in min- imum mean annual and winter temperatures, and by the intensification of seasonality.

(6) A further trend in the climate development along the consecutive dglacialsT is indicated by reduced minimum summer, mean annual, and winter temperatures.

(7) Considering maximum temperature shifts along the consecutive dinterglacial–glacialT transitions, the amplitudes are increasing for summer and mean annual temperatures. As compared to the amplitude of winter temperature change during the last dinterglacial–glacialT (Eemian–Weichse- lian) transition in South Italy, the Late Pliocene shifts are less than half in size.

Acknowledgements

S. Klotz and D. Uhl gratefully acknowledge the financial support by the Alexander von Humboldt- Foundation, and N. Combourieu-Nebout by CNRS for processing palynological samples (this is LSCE contri- bution No. 1804). S. Klotz gratefully acknowledges also the financial support by the Laboratoire Pale´oEn- vironnements and Pale´obioSphe`re, and the European Science Foundation. The authors thank Dr. L. Lourens, Prof. C. Tzedakis and an anonymous reviewer for their valuable comments on an earlier manuscript which helped to improve the manuscript.

References

[1] M.E. Raymo, W.F. Ruddiman, J. Backman, B.M. Clement, D.G.

Martinson, Late Pliocene variation in Northern Hemisphere ice sheets and North Atlantic deep circulation, Palaeoceanography 4 (1989) 413 – 446.

[2] N.J. Shackleton, New data on the evolution of Pliocene climatic variability, in: E.S. Vrba, G.H. Denton, T.C. Partridge (Eds.), Paleoclimate and Evolution, with Emphasis on Human Origins, Yale Univ. Press, New Haven, 1995, pp. 24 – 45.

[3] J. Zachos, M. Pagani, L. Sloan, E. Thomas, K. Billups, Trends, rhythms, and aberrations in global climate 65 Ma to present, Science 292 (2001) 686 – 693.

[4] W.H. Zagwijn, Subtropical relicts in the Pliocene flora of Bruns- sum (The Netherlands), Geol. Mijnb. 69 (1990) 219 – 225.

[5] N. Combourieu-Nebout, C. Vergnaud Grazzini, Late Pliocene northern hemisphere glaciation: the continental and marine responses in the central Mediterranean, Quat. Sci. Rev. 10 (1991) 319 – 334.

[6] J.P. Suc, Origin and evolution of the Mediterranean vegetation and climate in Europe, Nature 307 (1984) 429 – 432.

[7] C. Vergnaud Grazzini, J.F. Salle`ge, M.J. Urrutiaguer, A. Iannace, Oxygen and Carbon isotope stratigraphy of ODP hole 653A and site 654: the Pliocene–Pleistocene glacial history recorded in the Tyrrhenian Basin (West Mediterranean), Proc. Ocean Drill.

Prog., Sci. Results 107 (1990) 361 – 385.

[8] R.C. Thunell, D.F. Williams, E. Tappa, D. Rio, I. Raffi, Pliocene–

Pleistocene stable isotope record for ocean drilling program site 653, Tyrrhenian Basin, implications for the paleoenvironmental history of the Mediterranean Sea, Proc. Ocean Drill. Prog., Sci.

Results 107 (1990) 387 – 399.

[9] L.J. Lourens, F.J. Hilgen, L. Gudjonsson, W.J. Zachariasse, Late Pliocene to early Pleistocene astronomically forced sea surface productivity and temperature variations in the Mediterranean, Mar. Micropaleontol. 19 (1992) 49 – 78.

[10] L.J. Lourens, A. Antonarakou, F.J. Hilgen, A.A.M. Van Hoof, C.

Vergnaud-Grazzini, W.J. Zachariasse, Evaluation of the Plio- Pleistocene astronomical timescale, Paleoceanography 11 (1996) 391 – 413.

[11] D. Kroon, I. Alexander, M. Little, L.J. Lourens, A. Matthewson, A.H.F. Robertson, T. Sakamoto, Oxygen isotope and sapropel stratigraphy in the Eastern Mediterranean during the last 3.2 million years, Proc. Ocean Drill. Prog., Sci. Results 160 (1998) 181 – 189.

[12] W.H. Zagwijn, Aspects of the Pliocene and early Pleistocene vegetation in the Netherlands, Meded. - Rijks Geol. Dienst C 3 (1960) 1 – 78.

[13] J.-P. Suc, A. Bertini, N. Combourieu Nebout, F. Diniz, S. Leroy, E. Russo-Ermolli, Z. Zheng, E. Bessais, J. Ferrier, Structure of West Mediterranean vegetation and climate since 5.3 Ma, Acta Zool. Cracov. 38 (1995) 3 – 16.

[14] F.J. Hilgen, Astronomical calibration of Gauss to Matuyama sapropels in the Mediterranean and implications for the geomag- netic polarity time scale, Earth Planet. Sci. Lett. 104 (1991) 226 – 244.

[15] L.J. Lourens, Astronomical Forcing of Mediterranean Climate during the Last 5.3 Million Years, Proefschrift, Universiteit Utrecht, Utrecht, 1993, 247 pp.

[16] N. Combourieu-Nebout, Vegetation response to upper Pliocene glacial/interglacial cyclicity in the central Mediterranean, Quat.

Res. 40 (1993) 228 – 236.

[17] J. Iversen, Viscum, Hedera and Ilex as climate indicators, Geol.

Fo¨ren. Sto¨ckh. Forh. 66 (1944) 463 – 483.

[18] K.J. Sinka, T.C. Atkinson, A mutual climatic range method for reconstructing palaeoclimate from plant remains, J. Geol. Soc.

156 (1999) 381 – 396.

[19] S. Fauquette, J. Guiot, J.P. Suc, A method for climatic recon- struction of the Mediterranean Pliocene using pollen data, Palaeogeogr. Palaeoclimatol. Palaeoecol. 144 (1998) 183 – 201.

[20] S. Klotz, Neue Methoden der Klimarekonstruktion-angewendet auf quarta¨re Pollensequenzen der franzo¨sischen Alpen, Tu¨b.

Mikropala¨ontol. Mitt. 21 (1999) (169 pp.).

[21] S. Klotz, U. Mu¨ller, V. Mosbrugger, J.L. de Beaulieu, M. Reille, Eemian to early Wu¨rmian climate dynamics: history and pattern of changes in Central Europe, Palaeogeogr. Palaeoclimatol.

Palaeoecol. 211 (2004) 107 – 126.

[22] N. Combourieu-Nebout, Re´ponse de la ve´ge´tation de l’Italie me´ridionale au seuil climatique de la f in du Plioce`ne d’apres l’analyse pollinique haute re´solution de la section Semaforo (2,46 a` 2,1 Ma), C.R. Acad. Sci. Paris 321 (1995) 659 – 665.

[23] L.J. Lourens, F.J. Hilgen, N.J. Shackleton, J. Laskar, D. Wilson, The Neogene period, in: F.M. Gradstein, J.G. Ogg, A.G. Smith (Eds.), A Geologic Time Scale 2004, Cambridge University Press, 2004, pp. 409 – 440.

[24] G. Pasini, M.L. Colalongo, The Pleistocene boundary stratotype at Vrica, Italy, in: J.A. Van Couvering (Ed.), The Pleistocene Boundary and the Beginning of the Quaternary, Cambridge University Press, Cambridge, 1997, pp. 15 – 45.

[25] K. Faegri, J. Iversen, Textbook of Pollen Analysis, Hafner Press, New York, 1997, 295 pp..

[26] A.C. Mix, N.G. Pisias, W. Rugh, J. Wilson, A. Morey, T.K.

Hagelberg, Benthic foraminifer stable isotope record from Site 849 (0–5 Ma): local and global climate changes, Proc. Ocean Drill. Program Sci. Results 138 (1995) 371 – 412.

[27] J. Laskar, P. Robutel, F. Joutel, M. Gastineau, A. Correia, B. Levrand, A long-term numerical solution for the insola- tion quantities of the Earth, Astron. Astrophys. 428 (2004) 261 – 285.

[28] S. Klotz, J. Guiot, V. Mosbrugger, Continental European Eemian and early Wu¨rmian climate evolution: comparing signals using different quantitative reconstruction approaches based on pollen, Glob. Planet. Change 36 (2003) 277 – 294.

[29] J.-P. Suc, J. Cravatte, Etude palynologique du Plioce`ne de Catalogne (nord-est de l’Espagne), Pale´obiol. Cont. 13 (1982) 1 – 31.

[30] N. Combourieu-Nebout, S. Fauquette, P. Que´zel, What was the late Pliocene Mediterranean climate like: a preliminary quanti- fication from vegetation, Bull. Soc. Ge´ol. France 171 (2000) 271 – 277.

[31] W. Landini, E. Menesini, L’ittiofauna plio-pleistocenica della sezione delle Vrica (Crotone-Calabria), Bull. Soc. Palaeontol.

Ital. 17 (1978) 143 – 175.

[32] V.F. Ro¨gl, F.F. Steininger, Vom Zerfall der Tethys zu Mediterran und Paratethys. Die Neogene Pala¨ogeographie und Palinspastik des zirkummediterranean Raumes, Ann. Naturhist. Mus. Wien 85 (1983) 135 – 163.

[33] C.W. Wang, The Forests of China with a Survey of Grassland and Desert Vegetation, Maria Moors Cabot Foundation, Cam- bridge, Mass., 1961.

[34] L.E. Heusser, W.L. Balsam, Pollen distribution in the northwest Pacific Ocean, Quat. Res. 7 (1977) 45 – 62.

[35] J.L. Vankat, The Natural Vegetation of North America: An Introduction, Wiley, New York, 1979.

[36] S. Fauquette, J.-P. Suc, J. Guiot, F. Diniz, N. Feddi, Z. Zheng, E.

Bessais, A. Drivaliari, Climate and biomes in the West Medi- terranean area during the Pliocene, Palaeogeogr. Palaeoclimatol.

Palaeoecol. 152 (1999) 15 – 36.

[37] J. Pross, S. Klotz, Palaeotemperature calculations from the Praetiglian/Tiglian (Plio-Pleistocene) pollen record of Lieth, northern Germany: implications for the climatic evolution of NW Europe, Glob. Planet. Change 34 (2002) 253 – 267.

[38] M.E. Raymo, K.H. Nisancioglu, The 41 kyr world: Milanko- vitch’s other unsolved mystery, Paleoceanography 18 (2003), doi:10.1029/2002PA000791.

S. Klotz et al. / Earth and Planetary Science Letters 241 (2006) 174–187 186

[39] S. Leroy, L. Dupont, Development of vegetation and continental aridity in northwestern Africa during the Late Pliocene: the pollen record of ODP Site 658, Palaeogeogr. Palaeoclimatol.

Palaeoecol. 109 (1994) 295 – 316.

[40] K. McIntyre, M.L. Delaney, A.C. Ravelo, Millenial-scale cli- mate change and oceanic process in the late Pliocene and early Pleistocene, Paleoceanography 16 (2001) 535 – 543.

[41] M.W. Howell, R.C. Thunell, E. Di Stefano, R. Sprovieri, E.J.

Tappa, T. Sakamoto, Stable isotope chronology and paleo- ceanographic history of Sites 963 and 964, Eastern Mediter-

ranean Sea, Proc. Ocean Drill. Prog., Sci. Results 160 (1998) 167 – 180.

[42] F.J. Hilgen, L.J. Lourens, A. Berger, M.F. Loutre, Evaluation of the astronomically calibrated time scale for the late Pliocene and earliest Pleistocene, Paleoceanography 8 (1993) 549 – 565.

[43] J.R.M. Allen, W.A. Watts, B. Huntley, Weichselian palynostra- tigraphy, palaeovegetation and palaeoenvironment, the record from Lago Grande di Monticchio, Southern Italy, Quat. Int.

73/74 (2000) 91 – 110.